January 2004 - slime-devel - mailman3.common-lisp.net

[slime-devel] Daily ChangeLog diff
by Luke Gorrie 14 Jan '04

14 Jan '04

Index: slime/ChangeLog diff -u slime/ChangeLog:1.200 slime/ChangeLog:1.204 --- slime/ChangeLog:1.200 Mon Jan 12 23:23:38 2004 +++ slime/ChangeLog Wed Jan 14 01:54:11 2004 @@ -1,3 +1,58 @@ +2004-01-14 Luke Gorrie <luke(a)bluetail.com> + + * swank-openmcl.lisp (emacs-connected): Typo fix (missing + close-paren). + +2004-01-13 Helmut Eller <e9626484(a)stud3.tuwien.ac.at> + + * slime.el (slime-input-complete-p): Tolerate extra close parens. + (slime-idle-state): Don't active the repl. + (slime-insert-transcript-delimiter): Insert output before prompt. + (slime-open-stream-to-lisp): Initialize the process-buffer with + the connection buffer. + (slime-repl-activate): Deleted. + (slime-repl-eval-string, slime-repl-show-result) + (slime-repl-show-abort): Better handling of abortion. + (slime-compile-file): Insert output before prompt. + + * swank-lispworks.lisp (create-socket): Fix condition message. + + * swank-openmcl.lisp (*swank-in-background*): Set to :spawn. + (emacs-connected): Initialize ccl::*interactive-abort-process*. + + * swank.lisp (*swank-in-background*): New variable. + (start-server): Start swank in background, depending on + *swank-in-background*. + + * swank-cmucl.lisp, swank-sbcl.lisp (*swank-in-background*): Set + to :fd-handler. + + * swank-clisp.lisp (accept-connection): Remove superfluous call to + socket-wait. + + New more direct socket interface. The new interface is closer to + the functions provided by the implementations. For Lispworks we + use some non-exported functions to get a sane interface. The + interface also includes add-input-handler and a spawn function + (not used yet). The idea is that most of the logic can be shared + between similar backends. + + * swank-gray.lisp (make-fn-streams): New function. + (stream-read-char-no-hang, stream-read-char-will-hang-p): Moved to + here from swank-clisp.lisp. + + * swank-allegro.lisp, swank-clisp.lisp, swank-cmucl.lisp, + swank-lispworks.lisp, swank-openmcl.lisp, swank-sbcl.lisp: + (create-socket, local-port, close-socket, accept-connection) + (add-input-handler, spawn): Implement new socket interface. + + * swank.lisp (start-server, open-dedicated-output-stream &etc): + Use new socket functions. + + * swank-backend.lisp (create-socket, local-port, close-socket) + (accept-connection, add-input-handler, spawn): New functions. + (accept-socket/stream, accept-socket/run): Deleted. + 2004-01-13 Luke Gorrie <luke(a)bluetail.com> * swank-clisp.lisp: Updated for new network interface but not

1 0

[slime-devel] swank-openmcl
by Marco Baringer 14 Jan '04

14 Jan '04

that was a lot easier than i thought. -- Marco Ring the bells that still can ring. Forget the perfect offering. There is a crack in everything. That's how the light gets in. -Leonard Cohen

2 1

[slime-devel] swank-openmcl.lisp
by Marco Baringer 13 Jan '04

13 Jan '04

i'm in the process of porting the new backend to openmcl and i've managed to get emacs to block every single time i try to connect to slime. so far i've localized the problem to something inside slime-init-connection, but since emacs is blocked i can't get any debugging info (like the contents of stlime-events or access to inferior-lisp). whats the best way to go about debugging lime? -- Marco Ring the bells that still can ring. Forget the perfect offering. There is a crack in everything. That's how the light gets in. -Leonard Cohen

3 2

[slime-devel] Daily ChangeLog diff
by Luke Gorrie 13 Jan '04

13 Jan '04

Index: slime/ChangeLog diff -u slime/ChangeLog:1.199 slime/ChangeLog:1.200 --- slime/ChangeLog:1.199 Sun Jan 11 23:31:47 2004 +++ slime/ChangeLog Mon Jan 12 23:23:38 2004 @@ -1,3 +1,19 @@ +2004-01-13 Luke Gorrie <luke(a)bluetail.com> + + * swank-clisp.lisp: Updated for new network interface but not + tested! Probably slightly broken. + + * swank-lispworks.lisp: Updated for new network interface. + (accept-socket/stream): This function is currently broken, so + LispWorks can't use the dedicated output channel at the moment. + + * swank.lisp, swank-cmucl.lisp, swank-sbcl.lisp: Updated for new + network interface. + + * swank-backend.lisp (accept-socket/stream, accept-socket/run): + New functions replacing the ancient (over 24 hours!) + `create-socket-server'. This interface is much simpler. + 2004-01-12 Luke Gorrie <luke(a)bluetail.com> * swank-lispworks.lisp: Partially updated for new backend

1 0

[slime-devel] Problem in current version
by james＠unlambda.com 13 Jan '04

13 Jan '04

The current CVS version of SLIME is broken with XEmacs with SBCL. I get the following in the message area when I start SLIME: itimer "itimer-<1>" signaled: (invalid-argument "Marker does not point anywhere") And this is displayed in the *slime-repl* buffer: debugger invoked on a SB-INT:SIMPLE-PROGRAM-ERROR in thread 15191: invalid number of arguments: 2 restarts (invokable by number or by possibly-abbreviated name): 0: [ABORT ] Reduce debugger level (leaving debugger, returning to toplevel). 1: [TOPLEVEL] Restart at toplevel READ/EVAL/PRINT loop. ("XEP for EVAL-STRING" 2 "(swank:getpid)" NIL 1)[:EXTERNAL] source: ; file .../projects/lisp/slime/swank.lisp (DEFSLIMEFUN EVAL-STRING (STRING BUFFER-PACKAGE ID) (LET ((*PROCESSING-RPC* T) (*DEBUGGER-HOOK* #'SWANK-DEBUGGER-HOOK)) (LET (OK RESULT) (UNWIND-PROTECT (LET # # # # #) (SYNC-STATE-TO-EMACS) (FORCE-OUTPUT *SLIME-IO*) (SEND-TO-EMACS `#)))) (WHEN *DEBUGGER-HOOK-PASSBACK* (SETQ *DEBUGGER-HOOK* *DEBUGGER-HOOK-PASSBACK*) (SETQ *DEBUGGER-HOOK-PASSBACK* NIL))) 0] CL-USER> However, if I type into the *inferior-lisp* buffer things seem to be working still in SBCL. It was working fine a few days ago and I haven't done anything except update SLIME, so I presume it is due to a recent change. XEmacs 21.5.16, SBCL 0.8.7.10. Thanks, 'james -- James A. Crippen <james at unlambda.com> Lambda Unlimited 61.2204N, -149.8964W Recursion 'R' Us Anchorage, Alaska, USA, Earth Y = \f.(\x.f(xx))(\x.f(xx))

2 1

[slime-devel] Re: [slime-cvs] CVS update: slime/swank-clisp.lisp
by Luke Gorrie 13 Jan '04

13 Jan '04

Vladimir Sedach <vsedach(a)common-lisp.net> writes: > Update of /project/slime/cvsroot/slime > In directory common-lisp.net:/tmp/cvs-serv19072 > > Modified Files: > swank-clisp.lisp > Log Message: > Added support for the new backend. > > Date: Mon Jan 12 00:05:05 2004 > Author: vsedach Quick work, thanks! But please use the ChangeLog too so that people can keep up more easily. There're some details about the ChangeLog in the HACKING file. Cheers, Luke

1 0

[slime-devel] compile-file-for-emacs in swank-clisp
by Wolfgang Jenkner 12 Jan '04

12 Jan '04

> revision 1.6 > date: 2004/01/12 06:05:52; author: vsedach; state: Exp; lines: +29 -27 > File compilation and loading somewhat work now. CLISP's compile-file > returns some nutty number codes for warning-p and > failure-p. Meaningful compilation messages still on to-do list. Actually, I was convinced that my code was ;;; Ugly but essentially working. ;;; FIXME: I get all notes twice. (defmethod compile-file-for-emacs (filename load-p) [...] By the way, this includes compiler warnings and overlays at the correct lines. It seems that this didn't work for you. On the other hand, your patch does not work for me. So, perhaps, we should work this out somehow. Wolfgang

1 0

[slime-devel] Daily ChangeLog diff
by Luke Gorrie 12 Jan '04

12 Jan '04

Index: slime/ChangeLog diff -u slime/ChangeLog:1.196 slime/ChangeLog:1.199 --- slime/ChangeLog:1.196 Sat Jan 10 01:52:23 2004 +++ slime/ChangeLog Sun Jan 11 23:31:47 2004 @@ -1,3 +1,54 @@ +2004-01-12 Luke Gorrie <luke(a)bluetail.com> + + * swank-lispworks.lisp: Partially updated for new backend + interface, but not actually working. The sockets code is broken, I + haven't grokked LispWorks the interface properly. + + * swank-gray.lisp (slime-input-stream, slime-output-buffer): Added + slots to support the new `make-fn-streams' interface from + swank-backend.lisp. These slots need to be initialized by the + backend, see swank-sbcl.lisp for an example (very easy). + + * swank-sbcl.lisp (create-socket-server): Implemented new server + interface. + + * slime.el (slime-handle-oob): Added + :open-dedicated-output-stream message, previously implemented + with :%apply. + (slime-repl-read-string, slime-repl-return-string): Pass integer + argument to `slime-repl-read-mode' to set rather than toggle. + + * swank.lisp: Taking over previously non-portable jobs: + (start-server): Now only uses sockets code from the backend. + (handle-request): Top-level request loop. + (open-dedicated-output-stream): Dedicated output socket. + (connection): New data structure that bundles together the things + that constitute a connection to Emacs: socket-level stream and + user-level redirected streams. + + * swank-cmucl.lisp (create-socket-server): Generic TCP server + driven by SERVE-EVENT. + (serve-one-request, open-stream-to-emacs): Deleted. Now handled + portably in swank.lisp. + (make-fn-streams): Implement new stream-redirection interface. + (slime-input-stream): New slot referencing output sibling, so it + can be forced before input requests. + + * swank-backend.lisp (create-socket-server): Generic + callback-driven TCP server interface. Replaces + `create-swank-server', with the higher-level logic moved into + swank.lisp. + (emacs-connected): Invoked when Emacs initially connects, as a + hook for backend implementations. + (make-fn-streams): Interface for creating pairs of input/output + streams that are backended by callback functions. Used to + implement redirected-via-Emacs standard I/O streams. + +2004-01-12 Lawrence Mitchell <wence(a)gmx.li> + + * slime.el (slime-events-buffer): Set `hs-block-start-regexp' + before running `hs-minor-mode'. + 2004-01-10 Luke Gorrie <luke(a)bluetail.com> * slime.el (package-updating): Expected package is now a list (can

1 0

[slime-devel] Patch for (looking-at nil) bug in hideshow.
by lawrence mitchell 12 Jan '04

12 Jan '04

slime-events-buffer initialises hs-minor-mode before setting hs-block-start-regexp. This stops the *slime-events* buffer from being correctly set up, and subsequently stops slime from ever getting to the repl. The following patch fixes things: Index: ChangeLog =================================================================== RCS file: /project/slime/cvsroot/slime/ChangeLog,v retrieving revision 1.196 diff -u -r1.196 ChangeLog --- ChangeLog 10 Jan 2004 06:52:23 -0000 1.196 +++ ChangeLog 12 Jan 2004 00:03:59 -0000 @@ -1,3 +1,8 @@ +2004-01-12 Lawrence Mitchell <wence(a)gmx.li> + + * slime.el (slime-events-buffer): Set `hs-block-start-regexp' + before running `hs-minor-mode'. + 2004-01-10 Luke Gorrie <luke(a)bluetail.com> * slime.el (package-updating): Expected package is now a list (can Index: slime.el =================================================================== RCS file: /project/slime/cvsroot/slime/slime.el,v retrieving revision 1.172 diff -u -r1.172 slime.el --- slime.el 10 Jan 2004 06:43:52 -0000 1.172 +++ slime.el 12 Jan 2004 00:04:43 -0000 @@ -1484,8 +1484,8 @@ (let ((buffer (get-buffer-create "*slime-events*"))) (with-current-buffer buffer (lisp-mode) - (hs-minor-mode) (set (make-local-variable 'hs-block-start-regexp) "^(") + (hs-minor-mode) (setq font-lock-defaults nil) (current-buffer))))) -- lawrence mitchell <wence(a)gmx.li>

1 0

[slime-devel] Daily ChangeLog diff
by Luke Gorrie 10 Jan '04

10 Jan '04

Index: slime/ChangeLog diff -u slime/ChangeLog:1.190 slime/ChangeLog:1.196 --- slime/ChangeLog:1.190 Thu Jan 8 21:27:44 2004 +++ slime/ChangeLog Sat Jan 10 01:52:23 2004 @@ -1,3 +1,44 @@ +2004-01-10 Luke Gorrie <luke(a)bluetail.com> + + * slime.el (package-updating): Expected package is now a list (can + be any), since the shortest nickname is not + standardized. e.g. USER or CL-USER for COMMON-LISP-USER. + + * swank-cmucl.lisp: Don't enable xref (let the user decide). + (set-fd-non-blocking): Removed unused function. + Miscellaneous refactoring of the networking code. + + * slime.el (slime-complete-symbol): Use markers to hold the + beginning and end of the completion prefix, in case looking up + completions causes insertions (e.g. GC announcements). + +2004-01-09 Luke Gorrie <luke(a)bluetail.com> + + * slime.el (slime-activate-state): Only update state name when + `slime-default-connection' activates. This fixes an annoying + "Selecting deleted buffer" bug that prevented SLIME from being + restarted. + (slime-next-connection): Fixed a bug where buffer-local connection + bindings could get in the way and prevent the connection from + actually changing. + (slime-complete-restore-window-configuration): Wrap + `set-window-configuration' in `save-excursion'. This fixes a + problem where the cursor would end up in the wrong place after + completion in XEmacs. + +2004-01-09 Helmut Eller <e9626484(a)stud3.tuwien.ac.at> + + * slime.el: Place (require 'cl) inside a eval-and-compile. + (slime-with-connection-buffer): Move definition upwards before the + first use. + (package-updateing): New test for package updates in the listeners. + + * swank.lisp (eval-region): Bind *package* outside the + unwind-protect to detect updates. + + * swank-backend.lisp (debugger-info-for-emacs) + (find-function-locations): Doc fix. + 2004-01-09 Wolfgang Jenkner <wjenkner(a)inode.at> * swank-clisp.lisp: Add methods for GRAY:STREAM-READ-CHAR-NO-HANG

1 0

[slime-devel] annoying bug fixed: couldn't restart SLIME
by Luke Gorrie 09 Jan '04

09 Jan '04

Ahoy, Just checked in a fix for a really annoying problem where you couldn't restart SLIME having once started it ("selected deleted buffer" error message). Don't know how I missed it for a whole day, thanks Raymond Toy for pointing it out :-) -Luke

1 0

[slime-devel] (newbie) problem with packages under slime+cmucl (bug?)
by xur＠wp.pl 09 Jan '04

09 Jan '04

Hi. ---------------------------------------------------------------------- CMUCL: ---------------------------------------------------------------------- * lisp CMU Common Lisp CVS snapshot 2003-11, running on darkstar With core: /7/CMUCL/lib/cmucl/lib/lisp.core Dumped on: Sat, 2003-11-01 17:18:38+01:00 on lorien.users.earthlink.net See <http://www.cons.org/cmucl/> for support information. Loaded subsystems: Python 1.1, target Intel x86 CLOS based on Gerd's PCL 2003/06/18 09:23:09 * (setf *package* (make-package 'foo :use '(common-lisp))) #<The FOO package, 0/9 internal, 0/9 external> * *package* #<The FOO package, 0/9 internal, 0/9 external> ***-------------------- | ---------------------------------------------------------------------- | SLIME + CMUCL: | ---------------------------------------------------------------------- | CL-USER> (setf *package* (make-package 'foo :use '(common-lisp))) | #<The FOO package, 0/9 internal, 0/9 external> | | CL-USER> *package* | #<The COMMON-LISP-USER package, 2/9 internal, 0/9 external> ***------- ---------------------------------------------------------------------- regards, szymon.

3 3

Re: [slime-devel] Re: (newbie) problem with packages under slime+cmucl (bug?)
by Marco Baringer 09 Jan '04

09 Jan '04

[sorry for luke who is getting this twice] i found a possible bug in swank:eval-region where the package was being re-bound outside of the scope of test to see if the package had changed, the attached patch fixes this and makes the corrects the problem (for me at least). what i find strange is that if this "bug' really is a bug then how does slime work at all? Index: swank.lisp =================================================================== RCS file: /project/slime/cvsroot/slime/swank.lisp,v retrieving revision 1.88 diff -u -d -u -r1.88 swank.lisp --- swank.lisp 8 Jan 2004 07:02:20 -0000 1.88 +++ swank.lisp 9 Jan 2004 18:49:22 -0000 @@ -471,21 +471,21 @@ (force-output) (format-values-for-echo-area values))) -(defun eval-region (string &optional package-update-p) +(defun eval-region (string &optional (package-update-p t)) "Evaluate STRING and return the result. If PACKAGE-UPDATE-P is non-nil, and evaluation causes a package change, then send Emacs an update." - (unwind-protect - (do ((*package* *buffer-package*) - (str-length (length string)) - (pos 0) - (form nil) - (return-value nil (multiple-value-list (eval form)))) - ((= pos str-length) (values return-value form)) - (multiple-value-setq (form pos) - (read-from-string string nil nil :start pos))) - (when (and package-update-p (not (eq *package* *buffer-package*))) - (send-to-emacs (list :new-package (shortest-package-nickname *package*)))))) + (let ((*package* *buffer-package*)) + (unwind-protect + (do ((str-length (length string)) + (pos 0) + (form nil) + (return-value nil (multiple-value-list (eval form)))) + ((= pos str-length) (values return-value form)) + (multiple-value-setq (form pos) + (read-from-string string nil nil :start pos))) + (when (and package-update-p (not (eq *package* *buffer-package*))) + (send-to-emacs (list :new-package (shortest-package-nickname *package*))))))) (defun shortest-package-nickname (package) "Return the shortest nickname (or canonical name) of PACKAGE." -- Marco Ring the bells that still can ring. Forget the perfect offering. There is a crack in everything. That's how the light gets in. -Leonard Cohen -- Marco Ring the bells that still can ring. Forget the perfect offering. There is a crack in everything. That's how the light gets in. -Leonard Cohen

1 0

[Fwd: Re: [slime-devel] Re: (newbie) problem with packages under slime+cmucl (bug?)]
by xur＠wp.pl 09 Jan '04

09 Jan '04

1 0

[slime-devel] Daily ChangeLog diff
by Luke Gorrie 09 Jan '04

09 Jan '04

Index: slime/ChangeLog diff -u slime/ChangeLog:1.188 slime/ChangeLog:1.190 --- slime/ChangeLog:1.188 Wed Jan 7 13:10:43 2004 +++ slime/ChangeLog Thu Jan 8 21:27:44 2004 @@ -1,3 +1,16 @@ +2004-01-09 Wolfgang Jenkner <wjenkner(a)inode.at> + + * swank-clisp.lisp: Add methods for GRAY:STREAM-READ-CHAR-NO-HANG + and for the CLISP specific GRAY:STREAM-READ-CHAR-WILL-HANG-P. + This should fix the behaviour of SYS::READ-FORM. + +2004-01-08 Luke Gorrie <luke(a)bluetail.com> + + * slime.el (slime-inspector-fontify): Function to insert a string + in a particular inspector face. Replaces macro-code-generation + function `slime-inspector-expand-fontify'. Fixes a byte-compile + problem (macro was calling function not defined at compile-time). + 2004-01-07 Luke Gorrie <luke(a)bluetail.com> * slime.el: Multisession internal improvements. Now there are

1 0

[slime-devel] interrupt and multiple processes
by Alan Ruttenberg 08 Jan '04

08 Jan '04

When there are multiple processes and an interrupt is sent, lisp(openmcl in my case) needs to know which process to interrupt. So when the interrupt is being processed I need to ask emacs for that information before choosing the process to interrupt. What's the proper way to do that? Also regarding multiple processes: I'm running portable aserve and would like to be able to have errors that occur in the server processes create a slime debugger instance. Is there a way to set that up, given that those processes are not associated with a repl? Thanks, Alan

3 4

[slime-devel] possible typo in slime.el
by Raymond Wiker 08 Jan '04

08 Jan '04

My Emacs (GNU Emacs 21.3.50.1 / MacOSX) had a little trouble compiling the latest slime.el; it complained about the undefined function slime-inspector-expand-fontify. Examining this function, I thought it looked like it should be a macro instead of a function, so I changed "defun" to "defmacro". That seemed to work... is this the right fix?

3 4

[slime-devel] Daily ChangeLog diff
by Luke Gorrie 08 Jan '04

08 Jan '04

Index: slime/ChangeLog diff -u slime/ChangeLog:1.187 slime/ChangeLog:1.188 --- slime/ChangeLog:1.187 Tue Jan 6 20:21:22 2004 +++ slime/ChangeLog Wed Jan 7 13:10:43 2004 @@ -12,6 +12,8 @@ Fixed some related bugs. (slime-connection-close-hook): If default connection closes, select another connection. + (slime-lisp-package): Initially CL-USER nickname instead of + COMMON-LISP-USER (for REPL prompt). * slime.el (slime): Multisession support: with prefix argument, gives the option of keeping existing sessions and firing up an

1 0

[slime-devel] Assignment to free variable , real-var: each time I start Emacs
by Paolo Amoroso 07 Jan '04

07 Jan '04

After updating SLIME a couple of hours ago from the latest CVS sources, I get the following message in the *Compiler-Log* buffer each time I start Emacs: ** assignment to free variable ,real-var I use GNU Emacs 20.7.2 under Debian Woody. Paolo -- Why Lisp? http://alu.cliki.net/RtL%20Highlight%20Film

3 3

[slime-devel] Daily ChangeLog diff
by Luke Gorrie 07 Jan '04

07 Jan '04

Index: slime/ChangeLog diff -u slime/ChangeLog:1.184 slime/ChangeLog:1.187 --- slime/ChangeLog:1.184 Mon Jan 5 15:51:50 2004 +++ slime/ChangeLog Tue Jan 6 20:21:22 2004 @@ -1,3 +1,59 @@ +2004-01-07 Luke Gorrie <luke(a)bluetail.com> + + * slime.el: Multisession internal improvements. Now there are + three separate connection variables, in order of priority: + slime-dispatching-connection (dynamically-bound) + slime-buffer-connection (buffer-local) + slime-default-connection (global) + The most specific one available is used. This is splitting + `slime-connection' into multiple variables, so that you can be + specific about what you want to assign (i.e. know if you're + setting a dynamic binding or a buffer-local one). + Fixed some related bugs. + (slime-connection-close-hook): If default connection closes, + select another connection. + + * slime.el (slime): Multisession support: with prefix argument, + gives the option of keeping existing sessions and firing up an + additional *inferior-lisp* to connect to. Each connection now has + its own *slime-repl[<n>]* buffer. + (slime-connection): Should now be read via the function of the + same name. The accessor will check if the value is NIL, and if so + use `slime-default-connection'. + (slime-default-connection): The connection that will be used by + default, i.e. unless `slime-connection' is bound. Renamed from + `slime-primary-connection'. + (slime-init-connection-state): When reconnecting, update the + `slime-connection' binding in the REPL to use the new connection. + (slime-repl-input-history, ...): REPL variables are now + buffer-local. + +2004-01-06 Helmut Eller <e9626484(a)stud3.tuwien.ac.at> + + * swank.lisp (eval-string): New argument 'id'. Used to identify + the remote continuation. + (log-event): New debugging function. + (read-from-emacs, send-to-emacs): Use it. + + * slime.el: The new macro 'slime-rex' can now be used to evaluate + sexps remotely. It offers finer control what to do when the + evaluation aborts. + (slime-rex): New macro + (slime-eval, slime-eval-async, sldb-continue) + (sldb-invoke-restart): Use it. + (slime-continuation-counter, slime-push-evaluating-state): New + functions. + (slime-output-buffer): Initialize markers. + (sldb-mode): XEmacs doesn't like (add-hook (make-local-hook ...)). + (slime-init-connection): New optional argument SELECT. + (slime-def-connection-var): Workarounds for Emacs 20 reader bugs. + Backquote is pretty broken Emacs 20. + +2004-01-06 Ignas Mikalajunas <i.mikalajunas(a)mbt.lt> + + * swank-loader.lisp (user-init-file): Use merge-pathames. Fix + Windows support. + 2004-01-05 Luke Gorrie <luke(a)bluetail.com> * slime.el: Multiple session support, i.e. Emacs can open @@ -138,7 +194,6 @@ (slime-open-inspector): Minor indentation fixes. (slime-net-output-funcall): Removed. Was unused. ->>>>>>> 1.183 2003-12-19 Alan Ruttenberg <alanralanr(a)comcast.net> * slime.el 1.157 fix bug in sldb-princ-locals I introduced when adding fonts to sldb

1 0

[slime-devel] Bug report
by Ignas Mikalajunas 07 Jan '04

07 Jan '04

I am feeling lucky today ;) I was trying to profile one function that runs pretty long then got an error telling me that 589 ... ... ... is not a fixnum. After pressing enter on a [return to slime top level] a few times i got back to my lisp code. Then i tried to execute (profile:profile-all) (was compiling clis at that moment so PC was under a pretty heavy load) and got this (the report is attached). Ignas Mikalajunas

2 1

[slime-devel] Usable multiple session support
by Luke Gorrie 07 Jan '04

07 Jan '04

Howdyo, The multiple session support should be usable now. Each connection gets its own REPL, and if you do `C-u M-x slime' you get the option of firing up an extra *inferior-lisp* to connect to. Note that currently if you bring up two inferior-lisps, the first one will be called *inferior-lisp*<1> and the second *inferior-lisp*. As before, `slime-next-connection' will cycle the "default" connection that will be used for commands in your lisp-mode buffers. The REPL, SLDB, apropos-results, etc buffers are all "sticky" -- they keep using the connection that created them. Currently there isn't much to help you keep track of matters, except that repl/sldb buffers have the connection number in their buffer name. Obligatory screenshot (slightly gimp-mangled, I'm no Dan Barlow :-) http://www.bluetail.com/~luke/misc/lisp/slime-multisession.png Bug reports or usability suggestions would be welcome. Cheers, Luke

1 1

[slime-devel] Crash when invoking restarts in break loop view
by Paolo Amoroso 06 Jan '04

06 Jan '04

After updating SLIME half a dozen hours ago from the latest CVS head sources, and I have found a reproducible way of crashing it with CMUCL 18e and GNU Emacs 20.7.2 under Debian Woody. Do the following: 1) Create a file such as the following, which contains a Lisp form that generates an error: (defun toward-infinity () (/ 1 0)) (toward-infinity) 2) Evaluate each form with `C-x C-e' 3) When SLIME's break loop and backtrace view appears, press `q' or invoke another restart with the corresponding numer. The following message is displayed: ------------------------------------------------------------------------ You have encountered a bug in SLIME. The communication state machine received an event that was illegal for its current state, which means that the communication between Emacs and Lisp has lost synchronization. The connection to Lisp has therefore been closed. You can open a fresh connection with `M-x slime'. Please report this problem to your friendly neighbourhood SLIME hacker, or the mailing list at slime-devel(a)common-lisp.net. Please include in your report: A description of what you were doing when the problem occured, the version of SLIME, Emacs, and Lisp that you are using, the Lisp backtrace, if one was printed, and the information printed below: The event was: (:aborted) The state stack was: (slime-debugging-state slime-evaluating-state slime-idle-state) The content of the *slime-events* buffer: (:emacs-evaluate "(swank:getpid)" nil (:catch-tag slime-result-80998)) (activate) ; slime-evaluating-state (:%apply "slime-open-stream-to-lisp" (32806)) (:new-features ("PCL-STRUCTURES" "PORTABLE-COMMONLOOPS" "PCL" "PYTHON" "PENTIUM" "I486" "LINKAGE-TABLE" "MP" "GENCGC" "GLIBC2.1" "CMU18" "CMU18E" "RELATIVE-PACKAGE-NAMES" "CONSERVATIVE-FLOAT-TYPE" "RANDOM-MT19937" "HASH-NEW" "X86" "LINUX" "GLIBC2" "UNIX" "COMMON" "CMU" "NEW-COMPILER" "COMMON-LISP" "ANSI-CL" "IEEE-FLOATING-POINT")) (:ok 574) (activate) ; slime-idle-state (:emacs-evaluate "(swank:interactive-eval \"(defun toward-infinity ()\n (/ 1 0))\n\")" nil (:function (lambda (&rest --cl-rest--) (apply '(lambda (G81043 value) (save-current-buffer (set-buffer (symbol-value G81043)) (slime-show-evaluation-result value))) '--buffer-- --cl-rest--)))) (activate) ; slime-evaluating-state (:ok "TOWARD-INFINITY") (activate) ; slime-idle-state (:emacs-evaluate "(swank:interactive-eval \"(toward-infinity)\")" nil (:function (lambda (&rest --cl-rest--) (apply '(lambda (G81081 value) (save-current-buffer (set-buffer (symbol-value G81081)) (slime-show-evaluation-result value))) '--buffer-- --cl-rest--)))) (activate) ; slime-evaluating-state (:debug 1 ("Arithmetic error DIVISION-BY-ZERO signalled.\nOperation was KERNEL::DIVISION, operands (1 0)." " [Condition of type DIVISION-BY-ZERO]") (("ABORT" "Return to Slime toplevel.") ("ABORT" "Return to Top-Level.")) ((0 "(KERNEL::INTEGER-/-INTEGER 1 0)") (1 "(/ 1 0)") (2 "(SWANK:INTERACTIVE-EVAL \"(toward-infinity)\")") (3 "(SWANK:INTERACTIVE-EVAL 1 \"(toward-infinity)\")[:EXTERNAL]") (4 "(SWANK:EVAL-STRING \"(swank:interactive-eval \\\"(toward-infinity)\\\")\" NIL)") (5 "(SWANK:EVAL-STRING 2 \"(swank:interactive-eval \\\"(toward-infinity)\\\")\" NIL)[:EXTERNAL]") (6 "(SWANK::CALL-WITH-SLIME-STREAMS #<SWANK::SLIME-INPUT-STREAM {484CB23D}> #<Stream for descriptor 8> #<Two-Way Stream, Input = #<SWANK::SLIME-INPUT-STREAM {484CB23D}>, Output = #<Stream for descriptor 8>> #<Function FUNCALL {10060F79}> ...)") (7 "(SWANK::CALL-WITH-SLIME-STREAMS 5 #<SWANK::SLIME-INPUT-STREAM {484CB23D}> #<Stream for descriptor 8> #<Two-Way Stream, Input = #<SWANK::SLIME-INPUT-STREAM {484CB23D}>, Output = #<Stream for descriptor 8>> ...)[:EXTERNAL]") (8 "(SWANK::READ-FROM-EMACS)") (9 "(SWANK::READ-FROM-EMACS 0)[:EXTERNAL]") (10 "(SWANK::SERVE-ONE-REQUEST #<Stream for descriptor 6> #<Stream for descriptor 8> #<SWANK::SLIME-INPUT-STREAM {484CB23D}> #<Two-Way Stream, Input = #<SWANK::SLIME-INPUT-STREAM {484CB23D}>, Output = #<Stream for descriptor 8>>)") (11 "(SWANK::SERVE-ONE-REQUEST 4 #<Stream for descriptor 6> #<Stream for descriptor 8> #<SWANK::SLIME-INPUT-STREAM {484CB23D}> ...)[:EXTERNAL]") (12 "(\"DEFUN REQUEST-LOOP\" #<unused-arg>)") (13 "(\"DEFUN REQUEST-LOOP\" 1 #<unused-arg>)[:EXTERNAL]") (14 "(COMMON-LISP::SUB-SERVE-EVENT 1 0)") (15 "(SYSTEM:WAIT-UNTIL-FD-USABLE 0 :INPUT NIL)") (16 "(COMMON-LISP::DO-INPUT #<Stream for Standard Input>)") (17 "(COMMON-LISP::INPUT-CHARACTER #<Stream for Standard Input> NIL (COMMON-LISP::*EOF*))") (18 "(COMMON-LISP::SYNONYM-IN #<Synonym Stream to SYSTEM:*STDIN*> NIL (COMMON-LISP::*EOF*))") (19 "(COMMON-LISP::TWO-WAY-IN #<Two-Way Stream, Input = #<Synonym Stream to SYSTEM:*STDIN*>, Output = #<Synonym Stream to SYSTEM:*STDOUT*>> NIL (COMMON-LISP::*EOF*))"))) (activate) ; slime-debugging-state (:emacs-evaluate "(swank:throw-to-toplevel)" nil (:function (lambda (_)))) (activate) ; slime-evaluating-state (:aborted) (activate) ; slime-debugging-state (:debug-return 1) (:aborted) The content of the *cl-connection* buffer: ------------------------------------------------------------------------ The *inferior-lisp* buffer contains: * ;; Swank started at port: 32804. #<Handler for INPUT on descriptor 5: #<Closure Over Function "DEFUN ACCEPT-LOOP" {484C2D71}>> * ;; Connection to Emacs lost. ;; [End-of-File on #<Stream for descriptor 6>] After the error message is displayed, SLIME becomes unusable even after starting a fresh session. Incidentally, the error message has a typo: A description of what you were doing when the problem occured, ^^^^^^^ Paolo -- Why Lisp? http://alu.cliki.net/RtL%20Highlight%20Film

1 1

[slime-devel] sldb is not working in windows version of Allegro CL
by Ignas Mikalajunas 06 Jan '04

06 Jan '04

I have managed to track down the issue to these changes betwen two versions of swank-allegro.lisp (defmethod debugger-info-for-emacs (start end) - (list (debugger-condition-for-emacs) + (list (format-condition-for-emacs) (format-restarts-for-emacs) (backtrace start end))) (defmethod frame-locals (index) (let ((frame (nth-frame index))) (loop for i from 0 below (debugger:frame-number-vars frame) - collect (list :name (to-string (debugger:frame-var-name frame i)) + collect (list :symbol (debugger:frame-var-name frame i) :id 0 :value-string (to-string (debugger:frame-var-value frame i)))))) If i am not changing the file everything works perfectly yet after applying changes to the function frame-locals slime puts nil's instead of variable names. And modifying debugger-info-for-emacs effectively disables the debuger ... Ignas Mikalajunas

2 2

[slime-devel] Patch proposal
by Ignas Mikalajunas 06 Jan '04

06 Jan '04

To solve the portability issues (i am running slime in windows) i think the following source should be included into the slime distribution: --- swank-loader.lisp 2 Jan 2004 08:16:46 -0000 1.11 +++ swank-loader.lisp 5 Jan 2004 13:24:05 -0000 @@ -66,11 +66,10 @@ (defun user-init-file () "Return the name of the user init file or nil." (let ((home (user-homedir-pathname))) - (and (probe-file home) - (probe-file (format nil - #-mswindows "~A/.swank.lisp" - #+mswindows "~A\\_swank.lsp" - (namestring (truename home))))))) + (when (probe-file home) + (probe-file (merge-pathnames (truename home) + #-mswindows (make-pathname :name ".swank" :type "lisp") + #+mswindows (make-pathname :name "_swank" :type "lsp")))))) (compile-files-if-needed-serially (list* (make-swank-pathname "swank-backend") *swank-pathname* I am not sure whether the part with #-mswindows and #+mswindows is necessary windows at least win2K handles the .swank.lisp without any problems at all. The old version of this function was not working in windows anyway ;) i had to remove the "\\". Ignas Mikalajunas

2 1